Learning to Repair: Repairing model output errors after deployment using a dynamic memory of feedback
This addresses the issue of costly retraining for deployed models, offering a step towards more adaptable AI systems, though it is incremental as it builds on existing feedback mechanisms.
The paper tackles the problem of large language models making errors after deployment by enabling them to improve without retraining, using user feedback stored in a dynamic memory and a corrector model to repair outputs, resulting in up to 30 points improvement in error repair and up to 7 points improvement in avoiding similar mistakes.
Large language models (LMs), while powerful, are not immune to mistakes, but can be difficult to retrain. Our goal is for an LM to continue to improve after deployment, without retraining, using feedback from the user. Our approach pairs an LM with (i) a growing memory of cases where the user identified an output error and provided general feedback on how to correct it (ii) a corrector model, trained to translate this general feedback into specific edits to repair the model output. Given a new, unseen input, our model can then use feedback from similar, past cases to repair output errors that may occur. We instantiate our approach using an existing, fixed model for script generation, that takes a goal (e.g., "bake a cake") and generates a partially ordered sequence of actions to achieve that goal, sometimes containing errors. Our memory-enhanced system, FBNet, learns to apply user feedback to repair such errors (up to 30 points improvement), while making a start at avoiding similar past mistakes on new, unseen examples (up to 7 points improvement in a controlled setting). This is a first step towards strengthening deployed models, potentially broadening their utility. Our code and data is available at https://github.com/allenai/interscript/.